R/zfRenderSeurat.R

[HenrikBengtsson]

Splitting of from Comment #602 (comment) here:

Background

The code in R/zfRenderSeurat.R relies on functions that are not distributed via the seurat GitHub repository.

Issue

This makes it unnecessarily complicated to build the Seurat package from the GitHub repository - you need to (at least):

1. Drop R/zfRenderSeurat.R
2. Update the NAMESPACE file
3. Update the DESCRIPTION file

in order to not get flooded with R CMD check NOTEs about non-existing functions, specifically:

Undefined global functions or variables:
  hm4 open3d par3d play3d project.cell quads3d rotationMatrix spheres3d
  spin3d text3d view3d viewport

This needs to be repeated each time the package is updated upstream. This is turn makes it tedious to contribute.

Suggestions

Suggestion 1 - drop R/zfRenderSeurat.R

Distribute a clean version of Seurat without R/zfRenderSeurat.R, just as you do for CRAN. This can easily be done via separate branches and/or separate tags. For instance, before adding back R/zfRenderSeurat.R and then tagging with the version number, tag with the version number before such that the version number reflects exactly what's on CRAN, e.g. 2.3.4. Then, add back R/zfRenderSeurat.R and use a version number that is not clashing with what's on CRAN, e.g. 2.3.4-1000.

Suggestion 2 - provide / link to missing code

The R/zfRenderSeurat.R originates from the https://www.nature.com/articles/nbt.3192 publication. That article comes with Supplementary Code nbt.3192-S2.zip. However, none of the missing functions are defined there. Can these functions be found anywhere, or are they private functions?

[HenrikBengtsson]

Are open3d par3d play3d quads3d rotationMatrix spheres3d spin3d text3d view3d from the rgl package?

Is viewport from the grid package?

If so, that only leaves hm4() and project.cell() unknown.

We apologize for the hassle. As we are making changes to the package infrastructure, we will make sure its possible to contribute without these errors.

In the mean time, we would suggest that you perform steps 1-3 as you kindly outlined above, which should be a very quick solution, and hopefully not represent an undue burden.

[HenrikBengtsson]

Is it ok to send PRs where those have been dropped?

Of course